Data management platforms are used to integrate online and offline data to deliver a unified view of a group of third-party users. For example, the data management platform may collect large amounts of data describing characteristics or activities of third-party users that have interacted with different online services (e.g., social media websites, search engines, etc.) from various first-party activities. This data can describe, for example, network page visits, mobile application activity, offline activity, etc.
In some cases, operators of data management platforms or operators of online services may enter agreements with other parties (e.g., customers or other users of online services) regarding how the collected data will be used after its collection. For example, an operator of a data management platform may enter into contractual agreements with other parties (e.g., third-party providers) that define a set of privacy policies or other usage policies to which the operator must legally adhere. Thus, when disseminating the data to third-party providers or other destinations, the operator of a data management platform must comply with the agreed-upon restrictions on dissemination of the data.
In prior solutions for enforcing these restrictions, analysts are used to manually verify, prior to disseminating collected data, that the dissemination will not breach any contractual agreements. These solutions present disadvantages. For example, given the amount and complexity of the collected data, the likelihood of inadvertently violating a restriction is significantly increased when relying on manual review of the data. Furthermore, these solutions requiring manual review can increase the costs (e.g., time expenditure by analysts) of enforcing privacy restrictions or other usage policies. These solutions can also cause delays in delivering potentially sensitive data in response to requests from consumers of the data. Thus, prior solutions increase the cost of enforcing privacy restrictions or other usage policies while also increasing the risk that such restrictions will be accidentally violated, to the user. These disadvantages are exacerbated when the received data originates from multiple sources and has, potentially, multiple intended destinations.
It is therefore desirable to automatically enforce privacy restrictions or other usage policies on data received from multiple sources in a more accurate and efficient manner.